NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Green M, Wong M, Atkins D, et al. Diagnosis of Attention-Deficit/Hyperactivity Disorder. Rockville (MD): Agency for Health Care Policy and Research (US); 1999 Aug. (Technical Reviews, No. 3.)

Cover of Diagnosis of Attention-Deficit/Hyperactivity Disorder

Diagnosis of Attention-Deficit/Hyperactivity Disorder.

Show details

Summary of Findings

Question 1: Prevalence of ADHD in the General School-Age Population

The following prevalence studies of ADHD in the general population satisfied the inclusion criteria for this report: August and Garfinkel (1989); August, Realmuto, MacDonald, et al. (1996); Bird, Canino, Rubio-Stipec, et al. (1988); Cohen, Cohen, Kasen, et al. (1993); Costello, Costello, Edelbrock, et al. (1988); Costello, Edelbrock, Costello, et al. (1988); King and Young (1982); Kuperman, Johnson, Arndt, et al. (1996); Newcorn, Halperin, Schwartz, et al. (1994); Pelham, Gnagy, Greenslade, et al. (1992); Shaffer, Fisher, Dulcan, et al. (1996); Shekim, Kashani, Beck, et al. (1985); Tuthill (1996); Wolraich, Hannah, Pinnock, et al. (1996). Ten studies administered diagnostic instruments to representative samples of children identified in schools or in the general community (Table 3); three studies conducted an initial screening for certain symptoms among a random sample of children and administered diagnostic instruments only to those children who screened positive for symptoms (Table 7). Because of the difference in methodology, these studies are discussed separately.

Table 3. Selected ADHD Prevalence Data for Unscreened School-Age Population.

Table 3

Selected ADHD Prevalence Data for Unscreened School-Age Population.

Table 7. Estimated Prevalence for ADHD in Screened School-Age Population (across gender).

Table 7

Estimated Prevalence for ADHD in Screened School-Age Population (across gender).

These two groups of studies were analyzed separately. Two data files and analysis files were created. Only one study gave results using the DSM-IV criteria, and this study also gave results for DSM-III-R (Wolraich, Hannah, Pinnock, et al., 1996). Another study that provided prevalence rates using DSM-IV criteria is summarized narratively at the end of this section (Wolraich, Hannah, Baumgaertel, et al., 1998). Thus, the results were based on classification using the DSM-III or DSM-III-R classification scheme. Only one study gave results separately by single-year age categories. Most studies split the results at about age 10 (if they split by age all), and therefore 10 was used as the cut-point for age. Some studies split results by gender. When studies did not split the results by age or gender, they were assigned an age or gender fraction based on any information given about the study. No studies gave information separately by race, and this factor was not included in the analysis. The analysis table (Table 4) also includes the factors of setting (community or school) and diagnostic tool (DSM-III or DSM-III-R).

Table 4. Combined Estimates of Various Factors’ Effects on Unscreened School-Age Population Prevalence of ADHD.

Table 4

Combined Estimates of Various Factors’ Effects on Unscreened School-Age Population Prevalence of ADHD.

To examine whether age, gender, diagnostic tool, and setting influenced the estimated prevalence of ADHD, a multiple logistic regression model with random effects was used. This model explicitly recognizes that each study estimated ADHD rates under slightly different conditions (measurement method, population surveyed, informant, etc.). The analysis methodology is described by Hasselblad (1998). The analysis was done using the EGRET software from Cytel in Cambridge, MA. The results from the analysis of data from unscreened general population samples are presented in Table 4.

These results suggest that gender, diagnostic tool (DSM-III or DSM-III-R), and setting (community or school setting) are significant contributors to the prevalence of ADHD, but that age (5 to 9 years versus 10 to 12 years) is not a significant factor in this analysis.

These results give the answers in terms of odds rather than probabilities. The estimates were converted to probabilities and were adjusted so that the estimates represent the expected rates after 1986 (current practice). Using the results in Table 4, the rates of ADHD were estimated for the two gender groups as well as for study setting and criteria used. These estimates are presented in Table 5.

Table 5. Estimated Prevalence Rates of ADHD In School-Age Population by Gender, Setting, and DSM Version from Meta-Analysis.

Table 5

Estimated Prevalence Rates of ADHD In School-Age Population by Gender, Setting, and DSM Version from Meta-Analysis.

Only one study of the general population prevalence of DSM-IV ADHD was obtained for this report (Wolraich, Hannah, Baumgaertel, et al., 1998). Results from that study follow. A total of 214 elementary school (K-5) teachers in a Tennessee county consisting of 10 schools completed questionnaires rating each of 4,323 children on DSM-IV symptoms of disruptive behavior disorders, including attention-deficit/hyperactivity disorder, conduct disorder, and oppositional defiant disorder. In addition, ratings were obtained for each child’s level of impairment based on 10 items addressing academic and behavioral functioning. The questionnaires also included seven-item screenings for anxiety and depressive symptoms.

Results of the teachers’ ratings are presented in Table 6, which indicates the potential for much higher rates when impairment is not considered when making a diagnosis of ADHD. Table 6 suggests higher rates of ADHD across subtypes for boys versus girls, with the inattentive type being most common. The total rate when impairment is considered, 6.8 percent, does fall within the confidence interval of the analyzed studies that used earlier versions of the DSM.

Table 6. Prevalence of DSM-IV ADHD in School-Age Population According to Behavioral Impairment Criteria.

Table 6

Prevalence of DSM-IV ADHD in School-Age Population According to Behavioral Impairment Criteria.

Prevalence of ADHD in Screened General Population

Three studies utilized abbreviated screening procedures prior to diagnostic interviews for ADHD. Two of these studies (Bird, Canino, Rubio-Stipec, et al, 1988; Costello, Angold, Burns, et al., 1996) administered the diagnostic instrument to a subsample of “screen-negative” children as well as screen positive children to estimate how many ADHD cases may have been missed in the initial screen. In addition, these studies adjusted for loss of subjects between the initial screen and followup interviews. In contrast, the third study (August, Ostrander, Bloomquist, 1992) did not attempt to adjust for missed cases of ADHD among the screen-negative children (90% of total sample) or among screen-positive children who did not consent for diagnostic testing; it is thus likely to have underestimated the true prevalence of ADHD in its population. Data from these studies (Table 7) were analyzed separately from the single-stage population studies described earlier (Table 3). None of these studies presented gender-specific data.

Prevalence of Comorbid ADHD in General Population

Only four unscreened studies (August and Garfinkel, 1989; Pelham, Gnagy, Greenslade, et al., 1992; Shekim, Kashani, Beck, et al., 1985; Wolraich, Hannah, Pinnock, et al., 1996) and one screened study (Bird, Canino, Rubio-Stipec, et al., 1988) provided prevalence rates of various comorbid conditions in children with ADHD. August and Garfinkel (1989) gave results separately by age. Shekim, Kashani, Beck, et al. (1985) gave results separately by gender. Therefore, only the combined results (collapsed across age and gender) were analyzed. Rates (and 95 percent confidence limits) were computed for the comorbidities of oppositional defiant disorder, conduct disorder, anxiety disorder, depressive disorder, and learning disability. Rates for each of these were computed ignoring the presence of any other disorder. Multiple disorder rates were also summarized. Only the comorbidities of oppositional defiant disorder and conduct disorder were diagnosed in more multiple studies; anxiety and depression were diagnosed in two studies, and learning disorders in only one. These rates were combined using the empirical Bayes random effects model as described by Hedges and Olkin (1985, p. 199–200). The results are shown in Tables 8 through 13.

Table 8. Estimated Prevalence of Oppositional Defiant Disorder in Children With ADHD.

Table 8

Estimated Prevalence of Oppositional Defiant Disorder in Children With ADHD.

Table 13. Summary of Prevalence of Selected Comorbidities in Children With ADHD.

Table 13

Summary of Prevalence of Selected Comorbidities in Children With ADHD.

Results in Table 8 indicate that more than one-third of children diagnosed with ADHD also qualify for a diagnosis of oppositional defiant disorder.

Results in Table 9 indicate that more than one-quarter of children diagnosed with ADHD also qualify for a diagnosis of conduct disorder. Prevalence in the individual studies ranges widely, from 16 to 48 percent.

Table 9. Estimated Prevalence of Conduct Disorder in Children With ADHD.

Table 9

Estimated Prevalence of Conduct Disorder in Children With ADHD.

According to the results presented in Table 10, less than one-fifth of children with ADHD also have a depressive disorder.

Table 10. Estimated Prevalence of Depressive Disorder in Children With ADHD.

Table 10

Estimated Prevalence of Depressive Disorder in Children With ADHD.

Results in Table 11 suggest that more than one-quarter of children with ADHD qualify for a diagnosis of anxiety disorder.

Table 11. Estimated Prevalence of Anxiety Disorder in Children With ADHD.

Table 11

Estimated Prevalence of Anxiety Disorder in Children With ADHD.

Analyses of the results of the two studies in Table 12 suggest that almost one-third of children with ADHD also have more than one comorbid condition. Prevalence in the two studies ranges widely, from 16 to 50 percent.

Table 12. Estimated Prevalence of Multiple Comorbidities in Children With ADHD.

Table 12

Estimated Prevalence of Multiple Comorbidities in Children With ADHD.

Table 13 summarizes the prevalence of several comorbid conditions in children with ADHD, as discussed above.

Learning disabilities. Only one study used DSM criteria in examining the coexistence of learning disabilities in children with ADHD (August and Garfinkel 1989); this study estimated prevalence at 12 percent. Several other studies examining this issue were excluded because they used a dimensional measure rather than DSM-based structured diagnostic interviews. The reader may want to reference the following articles for a more comprehensive understanding of the relationship between ADHD and learning disorders: Brown, Madan-Swain, and Baldwin (1991); Robins (1992); Shaywitz, Shaywitz, Schnell, et al. (1988); and Stanford and Hynd (1994).

Overall, the prevalence of comorbid ADHD is high. Estimates of the prevalence of various comorbid conditions in children with ADHD range from 12 percent (learning disorders) to 35 percent (conduct disorder).

Only one study (Wolraich, Hannah, Baumgaertel, et al., 1998) provided rates of comorbid ADHD in the general population (see Table 14). The study also broke down the rates by subtype of ADHD. Rates appear to be consistent with previous versions of the DSM, according to results described earlier in this section, with high rates of comorbidity, particularly for oppositional defiant disorder and the anxiety or depressive disorders. Learning disability rates are correspondingly, and surprisingly, low. The low rate of learning disabilities in this study may reflect the lack of specific DSM criteria for learning disabilities. The main difference between the set of rates below and the rates discussed earlier in this section is a significantly lower rate of conduct disorder in the Wolraich, Hannah, Baumgaertel, et al., sample.

Table 14. Prevalence of Selected Comorbidities in Children With ADHD (DSM-IV).

Table 14

Prevalence of Selected Comorbidities in Children With ADHD (DSM-IV).

Summary of Question 1 Results

In the general, unscreened, school-age U.S. population, prevalence of ADHD ranged from 4 to 12 percent in studies using the DSM-III or DSM-III-R classification scheme. A multiple logistic regression analysis with random effects yielded results suggesting that gender, diagnostic tool, and setting are significant factors in the prevalence of ADHD, but that age is not significant. A single study using the DSM-IV classification scheme demonstrated that the prevalence of ADHD is substantially lower when impairment is required for the diagnosis than when impairment is not considered (7 percent vs. 16 percent). Higher rates of ADHD were found in boys than in girls for all types of ADHD, with the inattentive type most common.

In the general, unscreened, school-age population, prevalence of ADHD co-occurring with other disorders—oppositional defiant disorder, conduct disorder, anxiety disorder, depressive disorder, and learning disability—was estimated to be high, based on results of four studies. Of children diagnosed with ADHD, approximately 35 percent also qualified for a diagnosis of oppositional defiant disorder, 28 percent qualified for a diagnosis of conduct disorder, 26 percent qualified for a diagnosis of anxiety disorder, and 18 percent also had a depressive disorder. Learning disabilities in children with ADHD are estimated at a 12 percent prevalence.

Prevalence of ADHD in the screened school-age population was estimated at about 4 percent.

Question 2: Prevalence of ADHD in a Pediatric Clinic Setting

Two studies on the prevalence of ADHD in a pediatric clinic setting met inclusion criteria for this report. Because the results differed greatly, they were not formally analyzed. Instead, the findings from each study are summarized here.

In the first study (Lindgren, Wolraich, Stromquist, et al., 1989), primary care physicians identified 22 of 457 (4.8 percent) consecutive patients aged 6 to 12 years screened during the study period as positive for a “behavior disorder involving inattention and hyperactivity.” Of those 457 screened patients and an additional 10 who could not be seen during the screening period, 100 received a comprehensive evaluation. That set of 100 children included all 22 physician-identified children, 8 identified by their physicians as having another type of behavior disorder, 10 previously identified with ADHD by their physicians who could not be seen during the screening period, and a random sample of 60 of the screen-negative children. Comprehensive evaluations included DSM-III-R structured diagnostic interviews with parents, DSM-III-R-based checklists completed by teachers, and direct evaluation of the child with a continuous performance test, electronic activity monitoring, and examiner ratings of inattention, impulsivity, and hyperactivity. Results of the above modes of identifying children with ADHD are presented in Table 15.

Table 15. Effect of Diagnostic Criteria on Prevalence of ADHD in a Pediatric Clinic Setting.

Table 15

Effect of Diagnostic Criteria on Prevalence of ADHD in a Pediatric Clinic Setting.

Two articles present the results of the second study, each discussing different aspects of the study (Costello, Edelbrock, Costello, 1988; Costello, Costello, Edelbrock, et al., 1988). Children were sampled from a pool of 789 children who visited two HMO clinics in Pittsburgh between November 1984 and October 1985. The children were 7 to 11 years of age, primarily white (78 percent), middle (40 percent) to upper (39 percent) class, living in urban (42 percent) and suburban (58 percent) areas. Using the Child Behavior Checklist (CBCL), the authors screened for high-risk children, who were then evaluated for formal diagnoses. High risk was defined by a CBCL Total Problem Score above the 90th percentile of nonreferred children (T-score=70). Three hundred children were interviewed, including 126 children with high-risk CBCL scores and 174 with CBCL scores in the normal range. Each of these 300 children was interviewed separately.

First, a pediatrician interviewed the children and parents, employing criteria lists from the ICD-9A. Second, each child and parent was interviewed by a psychiatric social worker using the Diagnostic Interview Schedule for Children, Parent and Child Versions (DISC-P, DISC-C). Results indicated that three times as many children received ICD-9A diagnoses of “hyperactivity” from the pediatricians (n=12; weighted prevalence=1.5 percent) as DSM-III diagnoses of ADDH using the DISC (n=4; weighted prevalence=0.5 percent). Results are presented in Table 16.

Table 16. Prevalence of ADHD in a Pediatric Clinic Setting (in percentages) (Costello et al., 1988 study).

Table 16

Prevalence of ADHD in a Pediatric Clinic Setting (in percentages) (Costello et al., 1988 study).

The study also provided separate prevalence rates for DSM-III attention deficit disorder with hyperactivity (ADDH) and without hyperactivity (ADD). The rates differed depending on whether the informant was the parent (DISC-P) or child (DISC-C). Of the 300 children interviewed, 11 were diagnosed with ADDH using the DISC-P (weighted prevalence of 1.4 percent), and 5 were diagnosed with ADDH using the DISC-C (weighted prevalence of 0.6 percent). Two were diagnosed with ADD using the DISC-P (weighted prevalence 0.2 percent). No rate was provided for children diagnosed with ADD using the DISC-C. These results also are presented in Table 16.

Rates in the Costello et al. study are much smaller than in the Lindgren et al. study. Some of the difference may be attributable to the different versions of the DSM used; rates in studies have tended to increase with each new version of the DSM. However, the 1988 study’s rates are even smaller than expected for the earlier DSM. With only two studies, it is difficult to determine which is most accurate, but the Lindgren et al. study results are more consistent with rates found in general population studies.

Prevalence of Comorbid ADHD in Pediatric Clinic Setting

The same two studies were the only ones providing prevalence rates in a pediatric setting of ADHD with each of several comorbid conditions (Lindgren, Wolraich, Stromquist, et al., 1989; Costello, Costello, Edelbrock, et al., 1988).

In the Lindgren et al. study, comorbid conditions were diagnosed according to parent reports on the DISC-P, a structured diagnostic interview. Weighted estimates based on the total population of children seen during the study’s 3-month period (N=457) appear in Table 17. Of note, 59 percent of the children identified by physicians as having ADHD in this study had been placed in special-education classrooms because of learning disorders or developmental disorders, suggesting a relatively high rate of learning problems among these children.

Table 17. Prevalence of Selected Comorbid Conditions Among Pediatric Patients Diagnosed With ADHD.

Table 17

Prevalence of Selected Comorbid Conditions Among Pediatric Patients Diagnosed With ADHD.

In Costello, Costello, Edelbrock, et al. (1988), prevalence rates were calculated based on the structured diagnostic interviews with parents and children. Rates among DSM-III ADD or ADDH children ranged from 7.7 percent to 20 percent, depending on the comorbid condition and whether the informant was the parent (DISC-P) or child (DISC-C). Results are presented in Table 17. Again, rates in this study are much lower than those both in the Lindgren et al. study described above and in the general population. Changes in the DSM since 1988 cannot entirely explain the lower rates.

Summary of Question 2 Results

Two studies yielded information on prevalence of ADHD and comorbid ADHD diagnosed in a pediatric clinic setting. One study found that approximately 5 percent of children seen in a pediatric setting were diagnosed with ADHD and the other found fewer than 2 percent diagnosed with ADHD. Coexistence of ADHD with other disorders in children seen by a pediatrician was found in the first study to be 59 percent and in the other to range from 8 to 20 percent, depending on the comorbid condition and whether the informant was the parent or the child. These different results could not be explained, but the higher rates are consistent with rates found in general population studies.

Question 3: Reliability and Validity of Rating Scales for Diagnosis of ADHD

As discussed earlier, the scales were divided into two categories: ADHD-specific checklists and broad-band checklists. Behavior rating scales employ a series of questions (from 8 to more than 100) that parents or teachers answer regarding the behavior of the child. The responses are then converted into a numeric score. To determine how well a scale distinguishes children with ADHD from normal children, scores between the two populations can be compared.

Qualitative information on each of the scales is presented in Tables 18 and 19. The tables include information on the subscales included in each test, comorbid conditions addressed by each checklist, time required to administer, number of items, ages for which norms are available, computer scoring availability, and ordering information, including cost. Reliability and validity values are presented for each scale in Tables 20 and 21.

Table 18. Selected Behavior Rating Scales, Including Subscales, and Comorbid Conditions Addressed.

Table 18

Selected Behavior Rating Scales, Including Subscales, and Comorbid Conditions Addressed.

Table 19. Qualitative Information on the Behavior Rating Scales.

Table 19

Qualitative Information on the Behavior Rating Scales.

Table 20. Reliability and Validity of Various ADHD-Specific Behavior Rating Scales.

Table 20

Reliability and Validity of Various ADHD-Specific Behavior Rating Scales.

Table 21. Reliability and Validity of Various Broad-Band Behavior Rating Scales.

Table 21

Reliability and Validity of Various Broad-Band Behavior Rating Scales.

A common way to categorize the ability of psychological screening tests to discriminate abnormal from normal behavior is to calculate the “effect size,” based on the scores in case and control populations. The effect size is the difference in mean scores between two populations divided by an estimate of the individual standard deviation. The specific definition for the effect size of an experiment, d, used in this report is that given by Hedges and Olkin (1985). The effect size measure is easily interpreted. For example, an effect size of 4.0 means that the two populations are four standard deviation units apart and thus are almost completely separated. On the other hand, an effect size of 1.0 indicates much overlap of the two populations. Under some standard assumptions (see Hasselblad and Hedges, 1995), an effect size can be converted to a measure of sensitivity and specificity.

The columns in Table 22 are based on the concept that different cut-points can be used to vary the sensitivity and specificity. In clinical practice, definitions of an “abnormal” score on each rating scale are defined relative to the distribution of scores in a population of normal children; for most scales, an abnormal is one that is above the 90th or 95th percentile of scores in the normal control group (e.g. setting specificity of the scale at 90 or 95 percent). Table 22 illustrates the impact of effect size on the three sets of assumptions: (1) when the sensitivity and specificity are equal, (2) the effect on sensitivity when specificity is set at 90 percent, and (3) the effect on sensitivity when specificity is set at 95 percent. For readers who are familiar with ROC analyses, it is also possible to calculate the area under the curve directly from an effect size (assumptions given by Hasselblad and Hedges, 1995).

Table 22. Estimates of Sensitivity and Specificity for Various Values of the Effect Size.

Table 22

Estimates of Sensitivity and Specificity for Various Values of the Effect Size.

ADHD-Specific Checklists

Summaries of the results of effect size analyses for various ADHD-specific checklists used to screen for ADHD are shown in Tables 2326. The analyses reflected in these tables are based on studies done under ideal conditions. In these studies, children with ADHD were differentiated fairly well from normal controls (children presenting without significant problems). Such differentiation is not typical in actual practice, however, where children with ADHD often need to be distinguished from children presenting with a significant emotional or behavioral problem manifested in symptoms similar to those of ADHD. The actual performance of these tests in physicians’ offices with patients who have other conditions or comorbidities will be significantly poorer. The effectiveness of these measures in differentiating children with ADHD from children with other disorders could not be addressed in this report because the data were insufficient. More research is needed with these tools in discriminating children with ADHD from psychiatric controls.

Table 23. Total ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Table 23

Total ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Table 26. Impulsivity Subscales in ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Table 26

Impulsivity Subscales in ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Results in Table 23 indicate that the 1997 Revision of the Conners Rating Scale contains two highly effective indices for discriminating between children with ADHD and normal controls. The new ADHD Index and DSM-IV Symptoms Scale each achieved effect sizes greater than 3.0, which translate into matched sensitivity and specificity values of greater than 94 percent. Thus, use of these scales when comparing children with ADHD with normal controls will result in less than a 6 percent miss rate. Meanwhile, the effect sizes for the Barkley School Situations Questionnaire are weak, at less than 2.0, translating into matched sensitivity and specificity values of less than 86 percent. With each of these scales, however, only one study provided data; thus, interpretations of effectiveness must be made with caution.

As shown in Table 24, the hyperactivity subscales of several ADHD-specific checklists are strong in their ability to discriminate between children with ADHD and normal controls. The only hyperactivity subscale to perform poorly was from the ACTeRS Checklist. Its performance may have been lower in part due to a somewhat older sample; hyperactivity as a symptom tends to decrease with age. However, this factor alone would not account for the extremely low effect size. Again, several subscales are represented by only one study, requiring caution in interpretation of effectiveness.

Table 24. Hyperactivity Subscales of ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Table 24

Hyperactivity Subscales of ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Results presented in Tables 25 and 26 indicate that the SNAP Checklist’s inattention and impulsivity subscales discriminate well between children with ADHD and normal controls, with effect sizes greater than 4.0. Such performance can be translated into matched sensitivity and specificity values of greater than 97 percent. Meanwhile, the ACTeRS again performed poorly.

Table 25. Inattention Subscales of ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

Table 25

Inattention Subscales of ADHD-Specific Checklists: Ability to Detect ADHD vs. Normal Controls.

No data were found, even in the scales’ manuals, that compared mean performance of ADHD vs. normal controls with the following ADHD-specific checklists: Attention Deficit Disorders Evaluation Scale (ADDES), Barkley’s Home Situations Questionnaire (HSQ), Children’s Attention and Adjustment Survey (CAAS), Disruptive Behavior Disorders (DBD) Checklist, and DSM-IV Vanderbilt AD/HD Diagnostic Teacher Rating Scale (VADTRS).

Broad-Band Checklists

The purpose of this set of analyses is to determine which of the broad-band scales—those that screen for a variety of conditions including symptoms of ADHD—could serve as useful instruments in detecting clinical-level problems in children presenting at a pediatrician’s office. Any scale performing well could serve as a tool to screen for the many comorbid conditions typically found in children with ADHD. Unfortunately, the only data found compared the performance of referred versus nonreferred populations, rather than clinical versus normal populations. Because no diagnosis or screening was involved other than if a child had been referred for services, it is highly likely that many normals were among the referred and that many clinically significant problems were present in nonreferred children. Therefore, the results in this section should not be used to derive conclusions regarding the effectiveness of the scales in discriminating between children with significant problems versus children without significant problems.

Results of effect-size analyses for the broad-band checklists are described in this section. Data for these analyses compared mean performance of referred versus nonreferred populations. The bulk of the data used in this section were found in the scales’ published manuals (Achenbach 1991a,b,c; Burks, 1996; Conners, 1990; Lambert, Hartsough, and Sandoval, 1990; McCarney, 1995a,b; Naglieri, LeBuffe, and Pfeiffer, 1994; Reynolds and Kamphaus, 1992; Ullmann, Sleator, Sprague, et al., 1997).

The effectiveness of these scales’ global or total problem indices for discriminating referred from nonreferred populations is presented in Table 27. The analyses of the externalizing subscales, internalizing subscales, and then competence scales, are presented in Tables 28 through 30.

Table 27. Total Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 27

Total Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 28. Externalizing Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 28

Externalizing Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 29. Internalizing Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 29

Internalizing Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 30. Adaptive Functioning Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

Table 30

Adaptive Functioning Scales of Broad-Band Checklists: Ability to Detect Referred vs. Nonreferred.

The global or total scales are relatively consistent across the various studies, but the combined effect size of 1.5 represents a sensitivity and specificity of only about 80 percent. None of the tests had good estimated effect sizes for discriminating between referred and nonreferred populations.

The externalizing scales are less consistent across the various studies than are the total scales, but the combined effect size is similarly low (1.5). In general, this corresponds to a sensitivity and specificity of about 80 percent, which is not strong. None of the tests had a strong estimated effect size although the revised version of the Conners Parent Rating Scale appears to perform the best (see Table 28).

The internalizing scales are moderately consistent across the various studies, but the combined effect size of only 1.0 corresponds to a sensitivity and specificity of about 70 percent, which is poor. None of these tests had a good effect size for discriminating referred from nonreferred populations.

The adaptive functioning scales are very consistent across the various studies, probably because they come from the same parent scale (CBCL), in this analysis. Their combined effect size is low, at 1.2, corresponding to a sensitivity and specificity of about 72 percent, which is poor. Again, none of the tests had a good effect size.

Summary of Question 3 Results

Among ADHD-specific checklists, the 1997 revision of the Conners Rating Scale contains two highly effective indices for discriminating between children with ADHD and normal controls. The new ADHD Index and DSM-IV Symptoms Scale each are able to distinguish children with ADHD from normal controls 94 percent of the time. On the other hand, the Barkley School Situations Questionnaire is weak, with less than 86 percent effectiveness. Only one study provided data for these two tests and thus interpretations of effectiveness must be made with caution.

Hyperactivity subscales of the SNAP Checklist and the Conners Rating Scale are strong in their ability to discriminate between children with ADHD and normal controls. The only hyperactivity subscale to perform poorly was from the ACTeRS Checklist. The inattention and impulsivity subscales of the SNAP Checklist discriminated well between children with ADHD and normal controls, with effectiveness of greater than 97 percent. The ACTeRS Checklist performed poorly.

Broad-band checklists screen for a variety of conditions including symptoms of ADHD and serve as useful instruments in detecting the many comorbid conditions typically found in children with ADHD. Unfortunately, the studies reviewed could not be used to derive conclusions regarding the effectiveness of the scales in distinguishing children with significant problems from children without significant problems.

Question 4: Medical Screening Tests

A variety of different medical tests were proposed as part of the workup of children suspected of having ADHD. The purpose of the tests was to detect underlying causes or to help confirm a diagnosis by finding underlying abnormalities consistent with ADHD. This section examines studies trying to determine the likelihood of these tests diagnosing children with probable ADHD. Two categories of evidence were examined: (1) whether results of medical screening tests were significantly different in children with ADHD versus normal controls (e.g., mean TSH levels) and (2) how frequently screening tests detected conditions that required specific intervention (e.g., clinical hypo- or hyperthyroidism).

Lead Levels

Elevated levels of blood lead have been linked to a variety of adverse neurologic effects, ranging from symptomatic neurotoxicity at levels above 50 ug/dL to more subtle adverse effects on IQ and attention at milder elevations (10–50 ug/dL) in blood lead (U.S. Preventive Services Task Force 1996). The importance of elevated lead levels as a contributor to more severe behavioral problems, and the clinical value of blood lead measurements in the diagnosis of children with suspected ADHD remains controversial. A number of studies have suggested that children with increased lead levels in body tissues show decreases in cognitive ability, lower academic skills, and hyperactivity. Increased levels of lead have been reported in children with ADHD in some studies but not in others. Six studies were reviewed for this report (Barlow, 1983; Gittelman and Eskenazi, 1983; Kahn, Kelly, Walker, et al., 1995; Silva, Hughes, Williams, et al., 1988; Thomson, Raab, Hepburn, et al., 1989; Tuthill, 1996).

Table 31 presents the results of the analysis. Two studies showed no significant relationship between elevated lead levels and ADHD (Barlow, 1983; Gittelman and Eskenazi, 1983). One study did show such a relationship (Tuthill, 1996), and one study showed a weak association that did not reach statistical significance (Kahn, Kelly, Walker, et al., 1995). Two others demonstrated relationships between elevated lead levels and behavior problems (Silva, Hughes, Williams, et al., 1988; Thomson, Raab, Hepburn, et al., 1989). The study design, methods of linking lead levels to manifestations of ADHD, and sources and methods of lead level measurement differed across studies.

Table 31. Medical Screening Tests: Lead (Pb) Levels.

Table 31

Medical Screening Tests: Lead (Pb) Levels.

Although it appears possible that increased levels of lead play some role in ADHD, one can conclude that, overall, lead is not a major cause of ADHD. The dramatic decline in population lead levels in the U.S. over the past decade is likely to further reduce the role of lead as a contributor to attentional behavior problems, but we found only one study examining lead levels among a sample of ADHD patients in this country within the last 5 years. The available evidence suggests that routine lead screening contributes little to subsequent diagnostic or treatment strategies in children with suspected ADHD.

Thyroid

Abnormal thyroid function can have a range of behavioral effects, ranging from severe neuropsychological deficits in children with congenital hypothyroidism, hyperactivity associated with hyperthyroidism, and impaired concentration arising from hypothyroidism. For this report, four studies were reviewed for the relationship between abnormal thyroid function and ADHD (Elia, Gulotta, Rose, et al., 1994; Spencer, Biederman, Wilens, et al., 1995; Stein, Weiss, Refetoff, 1995; Weiss, Stein, Trommer, et al., 1993) (see Table 32). Data on TSH, or thyrotropin levels, were profiled, because this is the most routinely ordered thyroid level test. Not one study discovered a relationship between abnormal thyroid levels and ADHD. Overall, the prevalence of any thyroid disorder in children with ADHD appears to be the same as, or only slightly above, the prevalence of thyroid disorder in normal children. The evidence does not support the use of tests of thyroid function to screen for ADHD.

Table 32. Medical Screening Tests: Thyroid Levels.

Table 32

Medical Screening Tests: Thyroid Levels.

Interestingly, several of the studies from which data were pulled on TSH focused on the relationship between ADHD and a thyroid disorder called “Generalized Resistance to Thyroid Hormone” (GRTH). A high percentage of children with GRTH are diagnosed with ADHD as well. Researchers are intrigued by the relationship. Despite the fact that the relationship is unidirectional (children with ADHD rarely have GRTH), researchers are hoping it may shed light on the etiology of ordinary ADHD.

In the study by Elia, Gulotta, Rose, et al. (1994), 53 children with ADHD were screened for the presence of GRTH by several tests of thyroid function. No patient with ADHD was found to have GRTH. In a similar study by Spencer, Biederman, Wilens, et al. (1995), 132 children were examined for the presence of GRTH. Again, no patient with ADHD was found to have GRTH. In both of these studies, all thyroid hormone levels of the children with ADHD were in the normal range.

Weiss, Stein, Trommer, et al. (1993) studied 277 children with ADHD of whom none were found to have GRTH. However, 14 of 277 children with ADHD (5.4 percent) had some type of thyroid hormone abnormality, whereas only 1 of 106 normal children (1 percent) had such abnormalities. It should be noted that the abnormal levels in most of the children with ADHD in this study fell into the borderline range.

The last study, by Stein, Weiss, and Refetoff (1995) had a different focus. The authors studied 12 children with GRTH and 12 children with ADHD. The latter had completely normal thyroid function. The purpose was to compare the behavioral and cognitive characteristics of these two groups. The authors found that children with GRTH were similar behaviorally to children with ADHD, but differed in several other respects from children with the usual form of ADHD. Specifically, children with GRTH demonstrated lower nonverbal intelligence, or weaker perceptual-organizational skills, and lower academic achievement, suggesting more severe overall impairment in GRTH than in children with ADHD.

Imaging

A number of imaging studies of the brain have been performed to investigate whether any morphologic differences in various brain structures are present in children with ADHD. Morphologic differences might provide clues to biological correlates or causes of this disorder. A body of research exists on several biochemical and neurological pathways and processes known to mediate psychological and cognitive functioning within the brain. Gaining an understanding of structural or functional differences through imaging studies could lead to a more global understanding of the etiology of ADHD.

Nine imaging studies with children with ADHD were reviewed for this report (Table 33) (Castellanos, Giedd, Marsh, et al., 1996; Filipek, Semrud-Clikeman, Steingard, et al., 1997; Harcherik, Cohen, Ort, et al., 1985; Hynd, Semrud-Clikeman, Lorys, et al., 1990, 1991; Hynd, Hern, Novey, et al., 1993; Lyoo, Noam, Lee, et al., 1996; Semrud-Clikeman, Filipek, Biederman, et al., 1994; Shaywitz, Shaywitz, Byrne, et al., 1983). In two studies, no significant differences on brain CT or CAT scans were observed between children with ADHD and normal controls. In the other studies, several different abnormalities were noted in children with ADHD. Findings comprised either differences in size, in asymmetries, or in the shape or volume of the ventricles. In all cases, the structures in the children with ADHD were smaller than those of the normal control subjects.

Table 33. Medical Screening Tests: CT, CAT, and MRI.

Table 33

Medical Screening Tests: CT, CAT, and MRI.

In the future, a better understanding of ADHD is likely to evolve from the work currently being done with imaging. At the present time, however, the evidence is sparse and diverse. Therefore, none of these imaging procedures are supported by research as useful screening or diagnostic tools for ADHD.

Electroencephalography

One of the most researched medical tests used for evaluating children with ADHD is the electroencephalogram (EEG). This report abstracted data from eight studies seeking relationships between EEG patterns and ADHD (Holcomb, Ackerman, Dykman, 1985; Kuperman, Johnson, Arndt, et al., 1996; Lahat, Avital, Barr, et al., 1995; Matsuura, Okubo, Toru, et al., 1993; Newton, Oglesby, Ackerman, et al., 1994; Robaey, Breton, Dugas, et al., 1992; Satterfield, Schell, Nicholas, et al., 1990; Valdizan and Andreu, 1993) (see Table 34). None of the studies discovered any serious EEG abnormalities (e.g., signs of seizure activity) in children with ADHD. However, many found significant differences in brain wave activity between children with ADHD and normal controls.

Table 34. Medical Screening Tests: Electroencephalography.

Table 34

Medical Screening Tests: Electroencephalography.

Children with ADHD were found to have longer latencies at the P3 site (Holcomb, Ackerman, and Dykman, et al., 1985), longer latencies of certain waves for brainstem auditory evoked potentials (Lahat, Avital, Barr, et al., 1995), more slow waves and fewer Alpha waves (Matsuura, Okubo, Toru, et al., 1993), poorer ability to “tune” attention (Satterfield, Schell, Nicholas, et al., 1990), and asymmetry in peak amplitude evoked-response potentials (Kuperman, Johnson, Arndt, et al., 1996). The heterogeneity of results across studies prohibited meta-analysis and, as a result, indicates a lack of sufficient evidence of any clear EEG patterns typically found in children with ADHD. Therefore, evidence does not support routine use of the EEG as a screening tool for ADHD.

Neurological Screening Tests

In addition to the tests that try to relate anatomic structural differences and variations in biochemical neurotransmitter pathways to cognitive and behavioral function in children with ADHD, a number of other tests for neurological characteristics of children with ADHD have also been conducted, again to uncover clues to the etiology of ADHD. Studies are profiled in Table 35 (Accardo, Tomazic, Morrow, et al., 1991; Gillberg, Carlstrom, Rasmussen, et al., 1983; Reeves, Werry, Elkind, et al., 1987; Trommer, Hoeppner, Lorber, et al., 1988b; Vitiello, Stoff, Atkins, et al., 1990).

Table 35. Medical Screening Tests: Neurological Measures.

Table 35

Medical Screening Tests: Neurological Measures.

Minor anatomical malformations in various parts of the body have been associated with certain types of mental disorders. Accardo, Tomazic, Morrow, et al. (1991) examined 1,215 children presenting at a developmental center for various problems. The mean number of malformations in a group of 407 children with ADHD was no greater than the mean for a group of 808 controls.

The “Go-No-Go” test measures children’s ability to produce a simple motor response to a “go” cue and inhibit this motor response to the “no go” cue. Errors of omission, made when a child fails to “go” when cued, purportedly measure inattention. Errors of commission, or “going” when cued not to go, measure impulsivity. Trommer, Hoeppner, Lorber, et al. (1988b) administered this test to 16 children with DSM-III ADD, 28 children with DSM-III ADDH, and 32 controls. Both ADD and children with ADHD made more errors on this test than did controls. Moreover the ADHD groups made more omission errors (inattention) than did the ADD group. However, the authors report that error ranges were nearly identical in all groups, with absolute errors relatively low across groups. They state the test is not designed to be used diagnostically at this time.

Gillberg, Carlstrom, Rasmussen, et al. (1983) administered six neurological screening measures to several diagnostic groups, including an ADHD group and a normal control group. Across all measures, there were no significant differences in frequency of abnormal results between the ADHD and control groups. Specific frequencies for the groups were divided by gender. Half of the ADHD girls and none of the control girls displayed abnormal results on the Wechsler Intelligence Scale for Children subtest called “Mazes,” whereas all of the ADHD girls and half of the control girls obtained abnormal scores on the measure of diadochokinesis. Ten percent of the ADHD boys, versus 4–17 percent of control boys scored abnormally on four of the six tests. On the other two, measures of associated movements and diadochokinesis, about half of the ADHD boys scored abnormally, versus fewer than 20 percent of the control boys.

Reeves, Werry, Elkind, et al. (1987) examined the rate of neurodevelopmental abnormalities (as measured by nine tests of sensorimotor coordination) and minor physical anomalies (e.g., large head or low-set ears) in 39 ADDH and 39 control children. ADHD children demonstrated a significantly higher rate of neurodevelopmental abnormalities than matched controls, but no difference in prevalence of pre- or perinatal problems, speech problems, or minor physical abnormalities.

One study of the use of neurological screening tests with children screened for disruptive and impulsive behavior was profiled in Table 35 (Vitiello, Stoff, Atkins, et al., 1990). (In all other studies profiled in that table, children were diagnosed with ADHD.) In this screening study, the researchers administered the Revised Neurological Examination for Subtle Signs to 31 inpatient and outpatient children of a psychiatric clinic. Neurological soft, or subtle, signs did not correlate significantly with ratings of disruptive or impulsive behavior.

Miscellaneous Medical Tests

Because the cause of ADHD is unknown, because ADHD may not even be a single disease and may have several different causes, and because the complete workings of the brain and cognitive systems are still poorly understood, investigators have conducted a diverse array of studies searching for insights into the cause(s) of ADHD (Cacabelos, Albarran, Dieguez, et al., 1990; Cook, Stein, Ellison, et al., 1995; Hole, Lingjaerde, Morkrid, et al., 1988; LaHoste, Swanson, Wigal, et al., 1996; Pliszka, Maas, Javors, et al., 1994; Warren, Odell, Warren, et al., 1995). These include measurements of neurotransmitters, hormones, and proteins, including serotonin levels, plasma protein levels, peptide-containing urinary fractions, response to growth hormone releasing factor, dopamine receptors, and epinephrine levels. Studies for each are sparse, preventing formal statistical analyses to combine results of various studies or to determine trends or correlations between these measurements and ADHD. The studies are discussed below, however, and profiled in Table 36.

Table 36. Medical Screening Tests: Miscellaneous.

Table 36

Medical Screening Tests: Miscellaneous.

Cook, Stein, Ellison, et al. (1995) compared the following groups’ blood levels of the neurotransmitter, serotonin: ADHD with no comorbid conditions, ADHD with conduct disorder, ADHD with oppositional defiant disorder, and ADHD with both conduct disorder and oppositional defiant disorder. The latter group had a significantly higher rate of elevated serotonin levels than the other three groups.

Warren, Odell, Warren, et al. (1995) examined blood levels of the protein C4B in children with ADHD. C4B is a plasma protein that plays an important role in the body’s defenses against a variety of infectious agents. In this study, lower levels of C4B were found in children with ADHD than in normal controls. Mothers (but not fathers) of the children with ADHD also had significantly lower C4B levels compared with mothers of normal children. The authors noted a possible relationship between decreased concentrations of C4B and ADHD and speculated that, if replicated, low levels of C4B might represent a marker for ADHD. The decreased C4B levels in the mothers may have allowed a virus to persist during pregnancy, perhaps causing damage to the developing fetus, which then manifests itself after birth.

Hole, Lingjaerde, Morkrid, et al. (1988) investigated the patterns of peptide-containing urinary fractions of various sizes by gel filtration. All children with ADHD showed patterns atypical of those found in normal controls. However, these atypical patterns varied greatly across children with ADHD. Four groups of patterns were found among children with ADHD in this test; such heterogeneity precludes any possible correlations with ADHD at this time.

Cacabelos, Albarran, Dieguez, et al. (1990) studied the growth hormone response to growth hormone releasing factor (GRF). In 80–90 percent of children with ADHD, the growth hormone response was abnormal. Two different patterns of response were noted. Growth hormone is a neuropeptide that has certain known effects in the central nervous system and is part of the somatotropinergic system that could be involved in the etiology of ADHD.

LaHoste, Swanson, Wigal, et al. (1996) were interested in the role of the neurotransmitter, dopamine, and its receptor. Children with ADHD were found to have the less sensitive form of the dopamine receptor and, therefore would be expected to have reduced dopaminergic nerve impulse transmission. The authors suggest that the different variants of the gene coding for the dopamine receptor may be a factor in the expression of certain traits associated with ADHD.

Pliszka, Maas, Javors, et al. (1994) measured the urinary excretion of norepinephrine and epinephrine, or their metabolites, during a stressful task to evaluate the functioning of the noradrenergic system in children with ADHD. Regardless of the presence or absence of the comorbid condition, anxiety, children with ADHD excreted more normetanephrine (a metabolite of norepinephrine). Children with ADHD and anxiety excreted more EPI than did children with ADHD without anxiety. The authors conclude that, in some manner, EPI plays a role in the pathogenesis of ADHD.

In conclusion, the current evidence does not establish a relationship between any of the medical tests evaluated in this report and ADHD strong enough to warrant their use as routine screening or diagnostic tools in the evaluation of a child suspected of having ADHD.

Continuous Performance Tests (CPTs)

Data from the studies using CPTs were heterogeneous. Various types of CPTs were used with various scoring methods in studies using the CPT for many different purposes. The CPT studies that fit the inclusion criteria for this report are narratively outlined in Table 37 (August and Garfinkel, 1989; Barkley and Grodzinsky, 1994; Breen, 1989; Carter, Krener, Chaderjian, et al., 1995a; Cohen, Kelly, and Atkinson, 1989; Dykman and Ackerman, 1991; Fischer, Newby, and Gordon, 1995; Grant, Ilai, Nussbaum, et al., 1990; Halperin, Newcorn, Matier, et al., 1993; Horn, Wagner, and Ialongo, 1989; Loge, Staton, and Beatty, 1990; Seidel and Joschko, 1991). In addition to that qualitative analysis, a statistical analysis much like that performed with the ADHD-specific checklists was done with the CPT data. Specifically, the CPT scores for children diagnosed with ADHD were compared with those of normal controls to determine how effectively the CPT discriminates between the groups. Results are listed in Tables 3840, each of which presents the findings on one of the subtests within the CPT, including those that measure impulsivity, inattention/distraction, and vigilance. Only one study (Seidel and Joschko, 1991) provided data on total scales, so it is discussed in narrative at the end of this section.

Table 37. Continuous Performance Tests.

Table 37

Continuous Performance Tests.

Table 38. Impulsivity Measure in Continuous Performance Test (CPT): Ability to Detect ADHD vs. Normal Controls.

Table 38

Impulsivity Measure in Continuous Performance Test (CPT): Ability to Detect ADHD vs. Normal Controls.

Table 40. Vigilance Measure in Continuous Performance Test: Ability to Detect ADHD vs. Normal Controls.

Table 40

Vigilance Measure in Continuous Performance Test: Ability to Detect ADHD vs. Normal Controls.

The results listed in Table 38 show that the measures of impulsivity across various forms of the CPT are poor predictors of ADHD, with most effect sizes lower than 1.0. Corresponding sensitivity and specificity values would be less than 70 percent when those values are set to match.

Table 39 illustrates poor predictability of ADHD using the measures of inattention on various versions of the CPT. Effect sizes ranged from near 0 to just above 1.0, reflecting an inability of the measure to even distinguish the groups at one standard deviation from each other.

Table 39. Inattention Measure in Continuous Performance Test: Ability to Detect ADHD vs. Normal Controls.

Table 39

Inattention Measure in Continuous Performance Test: Ability to Detect ADHD vs. Normal Controls.

The only study using total scales was that of Seidel and Joschko (1991): the effect size for differences between total scores of ADHD and normal control males and females ages 6–11 was 1.147 with a 95 percent confidence interval of 0.509 to 1.784, again quite poor. Results for the total scale, as well as for vigilance measures (Table 40) are similar to those in Tables 38 and 39, indicating all subtests of the CPT are poor predictors of ADHD and would not serve as useful screening tools for ADHD, even when compared against normal controls, the most ideal of conditions.

Summary of Question 4 Results

A variety of medical tests have been examined as a way of detecting causes or abnormalities specifically associated with ADHD. Studies of tests for lead levels, abnormal thyroid function, morphologic differences in brain structures, EEG abnormalities (e.g., signs of seizure activity), and neurological characteristics were reviewed for this report. The studies sought to determine the likelihood of these tests to diagnose children with probable ADHD.

The ability of any of these tests to demonstrate a relationship to ADHD was not established. Significant lead levels were not found useful as a general tool for ADHD diagnosis. Studies describing the relationship of elevated lead levels and ADHD differed in their results. Overall, lead is not thought to be a major cause of ADHD, a conclusion strengthened by the fact that ADHD prevalence appears to be increasing, whereas lead levels in the population appear to be decreasing. Not one study discovered a relationship between abnormal thyroid levels and ADHD. Overall, the prevalence of any thyroid disorder in children with ADHD appeared to be the same as, or only slightly above, the prevalence of thyroid disorder in normal children. The evidence does not support the routine use of tests of thyroid function to detect underlying causes of ADHD. Although research on the thyroid disorder “Generalized Resistance to Thyroid Hormone” (GRTH) may shed light on fundamental mechanisms underlying ADHD, the rare nature of GRTH does not justify screening for it among children suspected of having ADHD.

Study results varied in regard to significant differences on brain computer tomographic (CT) or computed axial tomographic (CAT) scans between children with ADHD and normal controls. Two studies showed no significant differences, but several others noted abnormalities in children with ADHD, including differences in size, in asymmetries, or in the shape or volume of the ventricles. In all cases, the structures in the children with ADHD were smaller than those of the normal control subjects. A better understanding of ADHD is likely to evolve from work currently being done with imaging. At the present time, however, the evidence is sparse and diverse. Therefore, none of the imaging procedures are supported by research as useful screening or diagnostic tools for ADHD. No studies discovered any serious EEG abnormalities in children with ADHD. However, many found significant differences in brain wave activity between children with ADHD and normal controls. Overall, however, there was a lack of sufficient evidence of EEG patterns typically found in children with ADHD. Therefore, evidence does not support routine use of the EEG as a screening tool for ADHD. Finally, no studies found a significant correlation between neurological anomalies and ADHD.

This report reviewed the small number of studies of miscellaneous other tests to measure neurotransmitters, hormones, and proteins, including serotonin levels, plasma protein levels, peptide-containing urinary fractions, response to growth hormone releasing factor, dopamine receptors, and epinephrine levels. Trends or correlations between these measurements and ADHD could not be determined.

Finally, results on continuous performance tests and their subtests indicated poor prediction of ADHD. These tests would not serve as useful screening tools for ADHD.

PubReader format: click here to try

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...